Building the Ultimate Solid State Tesla Coil (Music-Capable!) | a Complete Guide

What's the ultimate, mad-science Halloween prop? Obviously the Tesla coil! These wonderful little devices are a staple of films like Frankenstein, and can produce continuous bolts of lightning at the throw of a switch! Unfortunately, some can be kinda tricky for beginners to make. But...what if I told you that you could build a single, customizable circuit for under $100 that can act like everything from a Van De Graaff generator to a vacuum tube coil to a full-fledged DRSSTC (all different kinds of high voltage devices with different spark appearances)? Sounds great, right? Well…you can!

Be sure to check out the official video on it here:

https://www.youtube.com/watch?v=zCf-PwXsG_E

This awesome new circuit was first designed by high voltage mastermind Steve Ward, modified by Gao Guangyan from loneoceans.com, and further optimized by me and my team of Tesla coil enthusiasts and experts, including:

• Max ("Power Max" on YouTube)
• Brian ("SciTubeHD" on YouTube)
• Jerry ("Arc Angel Tesla Coil" on YouTube)
• Ryan ("Magneticitist" on YouTube)
• Nick ("Coil Labs" on YouTube)
• Clayton ("ThePlasmaPrince" on YouTube)
• Walden ("Nerdlabs" on YouTube)

This circuit is solid state, interrupted, ran by an adaptable PCB, and can drive almost any coil to produce sparks over a foot long (30cm). And unlike the old LabCoatz circuit, this one can be ran for long periods of time without issue. No pesky ballast required! Best of all: this bad boy is also musical.

You will need the following parts:

Main SSTC circuit

- Two high-power IGBTs (FGA60N65SMD, also available from Digikey if Mouser stock is low!)

- One UCC27425 gate driver IC (UCC27425P or UCC27425PE4)

- One 74HC14 hex inverter Schmitt trigger (SN74HC14AN, TC74HC14APF, or SN74HC14N)

- One 555 timer (NE555P)

- One 7812 12V linear voltage regulator(L7812CV)

- One 7805 5V linear voltage regulator (L7805CV)

- Four 1N4148 diodes

- Two 1N4007 diodes

- One25V/220uF electrolytic capacitor(must have 3.5mm lead spacing to fit PCB and be 25V or greater)

- Two25V/470uF electrolytic capacitors(must have 5mm lead spacing to fit PCB and be 25V or greater)

- One5K resistor

- Two6.8 ohm resistors

- One 50K resistor

- One2.2K resistor

- One1K resistor

- One 2M potentiometer

- One 50K potentiometer

- One ON/OFF switch

- One 1uF ceramic capacitor (50V recommended)

- Two0.1uF film capacitor (10mm lead spacing, rated over 50V)

- One 0.33uF ceramic capacitor

- One 10uF ceramic capacitor (5mm lead spacing)

- Two0.82uF film capacitors(must have lead spacing between 13mm and 16mm, voltage over 450VDC, and a capacitance above 0.5uF)

- One toroidal ferrite(most of these are also suitable)

- TwoTO-247 compatible heatsinks

- Optional: one DIP-14 IC socket (DIP-14)

- Optional: two DIP-8 IC sockets(DIP-8)

- Optional: three terminal blocks (TB007-508-02BE)

Low voltage DC power supply

- One 12VAC transformer(this model is compatible with input voltages between 110V and 240V)

- One bridge rectifier (most units that can handle over 25V and a few amps will work fine)

High voltage DC power supply

- Two250V electrolytic capacitors (over 500uF)

- One SL32 1 ohm thermistor (inrush current limiter)

- One bridge rectifier (minimum recommended current/voltage ratings of 10A and 1000V)

- Optional: one ON/OFF switch (more information below)

Total price: $70 (this includes shipping)

Other materials needed:

- Wire (including magnet wire for the secondary coil), pipe (for secondary coil form), and miscellaneous hardware parts

- Soldering iron and solder

- Either the official PCB (details below) or breadboard

- Optional but useful: a handheld multimeter, oscilloscope, and Variac

Listed above is every part needed to build a working half-bridge solid state Tesla coil. However, please understand that the links provided may eventually expire. If a part is unavailable, you are encouraged to look up and find a similar replacement. Most of these part values are not super critical, and substitutions can be easily made. Additionally, if you possess parts of similar values, feel free to save money and use your parts instead! Owning a few large electrolytic capacitors and a 14V power source could save you an extra $20. With all the parts I had, I actually was able to build my SSTC for under $50!

These parts may also be found on other sites, such as digikey.com, newark.com, and even Amazon, so feel free to hunt for better deals!

It is also worth noting that having a few extra components on hand isn’t a bad idea. The components to worry about include the IGBTs, the UCC27425, the 74HC14, the 555 timer, and the high voltage supply’s bridge rectifier. The IGBTs are the most likely item to break, followed by the bridge rectifier (the rectifier typically dies from the overcurrent caused by IGBT failure). Minor screw-ups can also lead to the IC chips blowing, and they are dirt cheap, so it’s smart to buy an extra or two of each.

After you’ve settled on which parts to order, it’s time to start thinking about what you’ll mount them on. I personally recommend the official PCB that I designed for this circuit (see the video tutorial), but breadboard is also an option. The PCB should only cost a few dollars, is thoroughly labeled, and highly user-friendly. For the remainder of this instructional, I will be primarily talking about constructing the circuit on this PCB.

To order the PCB, simply take the “SSTC PCB” .zip file you can find here and submit it to a PCB manufacturing website like JLCPCB.com. This .zip file is technically referred to as a gerber file. If you want, you can change how many PCBs you order (minimum with JLCPCB is 5 copies for around $2 plus shipping), as well as the PCB color.

EXCLUSIVE! PLASMA CHANNEL PCB!

In a recent collaboration with Jay from Plasma Channel, I condensed my old version of this PCB quite a bit and made sure to add the power circuitry (the voltage doubler circuitry and rectifiers) so everything fits onto one PCB! You can find the PCB Gerber files for that here: https://drive.google.com/file/d/1gJerlIFWQG2c2BZqT3aXbNelYKJ1z2i8/view?usp=sharing

This is the circuit we will be building; it’s basically the Loneoceans SSTC 2 circuit with a few modifications. For one, I swapped the microcontroller-based interrupter with a more traditional 555-timer circuit, as it is far more user friendly and doesn’t require programming. I additionally changed and removed a few other components, including the undervoltage protection part of the circuit, after I found it was causing more harm than good. An NTC thermistor was also added to the main power section to prevent dangerously high inrush currents from destroying the rectifier and IGBTs. The need for this was actually brought up by Brian from SciTubeHD, who learned himself by blowing several components in his own copy of the Loneoceans SSTC 2.

How it works:

The driver for this coil works by first extracting a crude feedback signal from an antenna near the coil and passes it by a two-diode array. As Power Max demonstrated to me over a Zoom call, these two diodes essentially lock the positive and negative voltage peaks of the AC signal to a set value, preventing dangerously high voltages from entering the sensitive driver circuit.

Next, the signal passes through a resistor, capacitor, and 74HC14 Schmitt trigger, which converts the somewhat sloppy signal into a more functional squarewave that matches the resonant frequency. The new squarewave signal is then pumped into the UCC27425 gate driver IC, where it is amplified. Conveniently, our gate driver chip has something called an enable pin, which is basically like its on-and-off switch. If we feed the interrupter’s signal to the enable pins, we can control the Tesla coil’s pulse duration and frequency, and therefore the spark appearance.

The resulting interrupted signal from our driver circuit is finally sent to a small device known as a gate drive transformer, or GDT. Once properly assembled, the GDT converts the single 12V signal from our driver into two 18 volts signals, which are optimal for switching our transistors. If we phase the GDT correctly, it will cause the transistors to switch the DC voltage from our power supply across the primary coil at the resonant frequency. And since the resonant frequency is detected by the drive circuit, we can stick basically any secondary coil in the primary coil’s field and it will resonate almost perfectly, producing an extremely powerful electrical discharge.

The power sections:

To power this circuit, we need two power sections: one for low voltage and another for high voltage. All of the low-voltage circuitry is powered by a few simple voltage regulators, which take any DC voltage from 14 to 24 volts and adjust it and to fit the circuit’s needs. For our purposes, we can just take any 12VAC transformer and rectify it with a bridge rectifier to get around 17VDC.

For the high voltage input, this SSTC can take up to 400VDC. However, our wall sockets can only give us around 120 volts or 240 volts AC, depending on where you live. Fortunately, the circuit shown in the schematic takes care of this issue: with the flick of a switch it can either function as a full bridge rectifier (to get 340VDC in 240VAC regions or 170VDC in 120VAC regions) or a voltage doubler. The voltage doubler is only necessary for people in 110-120V regions, since doubled 240V (680VDC) would simply annihilate the SSTC circuit.

FOR THE PCB: the 340VDC supply and 14-24VDC supply are off-board (a design choice I made to save money on PCBs). For the low voltage supply, there is a port for the raw 14-24VDC input, and for the higher voltage, there are two positions to solder the power inputs to, as well as a ground port (marked ‘G’) to connect to mains ground. REMEMBER! Only direct current can enter this PCB! I've also provided a rough diagram to help wire up the PCB’s power and coils.

Gate drive transformer (GDT) winding:

For the gate drive transformer, you’ll want to wrap two twelve-turn coils and one eight turn coil onto a suitable ferrite core (more details about that in Step 3). Iron cores are almost non-functional at these high frequencies, so I advise against using them. As you can see in the circuit schematic, the two twelve turn coils are connected with “opposing polarity” (they are 180 degrees out-of-phase). This is necessary for the transistors to switch correctly. If you mess up the GDT phasing, the transistors will almost certainly die, so PAY ATTENTION TO HOW YOU WIND THEM! My PCB takes care of the phasing for you, and has marks to indicate phasing. If you have a hard time understanding this, check out the image included above!

One final note: if you’re making this circuit without the PCB, make sure to ground the low-voltage negative/neutral line. If you don’t, you’ll probably have tons of issues with interference, even with good shielding.

Perhaps one of the best parts about this circuit is its ability to handle changes and customization. As mentioned before, a number of it’s parts don’t need to be super exact, and can be changed for similar parts if need be. This section is your guide to customizing the designs and making them your own.

Necessary parameters:

These are the general requirements for each circuit component. If you follow what I say here, you should be good to go:

555 timer, 74HC14, and UCC27425: as far as I can tell, there are no suitable replacements for these parts. If you are more experienced in electronics design, feel free to look around though!

UPDATE FROM OLDER, MORE EXPERIENCED ZACH: the UCC27525 is a more powerful substitute for the UCC27425. Other UCC option might also exist, just look at the datasheets for similar pinouts!

Gate drive transformer (this is mostly regarding the ferrite itself):

• Initial permeability (μ): over 2000
• Diameter: around 20-30mm, it all depends on what you’re comfortable winding. Larger toroids are typically easier!
• Recommended materials: 77, 75, and N87 are best, but 43 is also usable
• Wire: most thin, insulated wire. DON’T USE MAGNET WIRE! Magnet wire will simply arc over, despite the low voltages.

Resistors:

• Power rating: most can be 1/4W (maybe less). The 6.8 ohm resistors should be around 1 watt though, since they’ll be handling more power.
• Resistance: try to stay within a few kiloohms of the recommended values, except for the 6.8 ohm resistors. For them, anything between 5-15 ohms should work.

Capacitors:

• Voltage: should be above 25V for the low voltage areas. For the large electrolytic capacitors in the doubler/rectifier, the voltage rating must be 250VDC or more. For the DC blocking capacitors in the power inverter section near the transistors, the voltage must be even higher: over 450V is needed, and over 500V is recommended. This is because they will be experiencing resonant voltage spikes, which can easily shred smaller 250V capacitors (ask me how I know).
• For the small electrolytic capacitors (used in the low voltage regulator section), any high capacitance (over 200uF) should work
• More than 500uF is recommended for the doubler capacitors
• For the small-value film and ceramic capacitors it’s best to stay within range of the ratings given. Variation by 75% or more are likely acceptable though.
• For the interrupter capacitor, you can honestly use any value you want. Just be aware that it will affect your coil’s BPS and pulse width. Lower values yield higher-BPS, thinner sparks, and higher values give thicker, low-BPS sparks. A good range to experiment in is 100pF to 100uF. I suggest using a value between 0.2uF to 0.5uF because it works well with the potentiometers to give the widest variety of spark appearances.

Diodes: most fast-recovery signal diodes will work

Bridge rectifiers:

• Reverse voltage: over 50V for the low voltage one (since it’s only handling around 25V max from the transformer) and around 1000V for the high voltage one
• Current: 4A or more for the low voltage one and 10A or more for the high voltage one. Peak current rating for the high voltage one should also be similar to or greater than that of the transistors (60A or so is usually fine)

Transistors:

• Type: either IGBT or N-channel MOSFET, and should have an internal diode. If no diode is built in, one must be added (look up info on flyback or freewheeling diodes)
• Power: over 250W (over 400W if you plan to operate at higher BPS/duty)
• Voltage: 500V or more
• Current: less important, but I recommend over 10A
• Peak current: most important, should be over 60A

Linear voltage regulators (7812 and 7805): most 12V and 5V positive linear voltage regulators that can output over 1A will work

Potentiometers:

• Power rating: 1/5W or more
• Resistance: anything you want, but, like the interrupter capacitor, it will affect your coil’s output. For the 2M potentiometer, using a lower value will increase the minimum BPS. For the 50K potentiometer, a lower value will decrease the maximum pulse width (resulting in thinner sparks)

Thermistor (inrush current limiter): should be an NTC thermistor rated for around 1 ohm and over 20A. I recommend looking at the SL32 line of thermistors, but feel free to explore other options!

For this circuit (and most IGBT-based Tesla coils), the resonant frequency shouldn’t exceed 400kHz. Above 400kHz, the IGBTs may still function, but they will do so with increasing difficulty and may wear down quicker. Since this coil is self-tuning, the resonant frequency is determined by the secondary coil and its topload. By putting the design parameters of your coil into a program like JavaTC, you can calculate the resonant frequency, as well as some other important values, such as coupling. With solid state Tesla coils like this, it best to leave the coupling below 0.4, with best results around 0.3. If the coil is over-coupled, it will tend to arc over to the primary coil, and if the coil is under-coupled, there will be less output.

One of the most critical aspect of your coil’s design is the primary impedance (this is just a fancy term for AC resistance). If the impedance is too low, excessive current will flow through the primary coil and kill your IGBTs, and if the impedance is too high, your sparks will be noticeably smaller. For the IGBTs I’m using, the primary impedance should be over 6 ohms in order to keep the peak currents within the IGBT’s rating.

Calculating impedance can actually be pretty simple: just go to a website like this that calculates LC impedance and input your coil’s specifications. The inductance is that of your primary coil and the frequency is just your secondary coil’s natural resonant frequency. Both of these values can be calculated using our old friend, JavaTC. For the capacitance, you’ll just use the value of one of our inverter capacitors. If you’re following my schematic, this value should be around 0.82 microfarad. Once you hit calculate, the value you’re interest in is the one marked ‘Total LC Impedance’. Once again, make sure you’re just over 6 ohms of impedance. In general, five turns on the primary should get you near this point.

While it’s totally possible for you to wire this circuit up and get it running well on the first try, it's also possible that you'll experience a few bumps. Fortunately, they are usually quite manageable and can be overcome with minimal effort. Here are a few of my best tips for troubleshooting this circuit:

If there is no output:

-     First, check all of your connections. I’ve been totally lost on several occasions, and the culprit turned out to just be a bad connection

-     Try bringing the feedback antenna closer to the coil. I’ve found my coil works best when the antenna is within a foot (30cm) or so of the secondary coil

-     Test your power sections and make sure they are getting (and giving!) power

-     If tons of current is being pulled, it’s possible you wired your GDT wrong and killed your IGBTs. You may have killed them in other ways, but this is the most common way. It’s also possible there is a random, accidental short somewhere in the circuit

-     Try swapping the primary coil connections.

If the output is very weak:

-     Try swapping the primary coil connections. This almost always solves the issue in this case.

-     The antenna might not be close enough, try moving it closer. A larger antenna is also helpful.

-     Check the interrupter on/off switch. I’ve had times when it was open (no interrupter signal being fed), and there was enough residual charge/interference to run the coil, albeit in a weaker state. This problem is usually indicated by a continuous, fuzzy plasma discharge (no buzzing or pulses).

-     Check the secondary coil’s ground connection. If a coil isn’t properly grounded, the voltage differential at the top could be lower with respect to ground, and therefore yield weaker output.

-     Your coupling could be very low or your primary impedance could be extremely high. Try running some calculations and see if anything jumps out as strange or off.

If you have unstable output (uneven BPS or pulse width)

-     This is usually a result of interference. Make sure your PCB’s ground connection is good, or, if you’re building it without the PCB, make sure the low voltage negative/neutral is grounded. Also, try shielding the circuit with a layer of grounded metal or metal foil.

Once the design, assembly, and troubleshooting are out of the way, it’s time to fire up the circuit and give it a run for its money! When built properly, this circuit can easily produce sparks larger than the secondary coil itself. With my unit, I’ve been able to achieve 12” sparks from an 8” coil, and 14” sparks with a similarly-size coil. These results are comparable to many of the commercial Tesla coils you can buy. The only difference: this coil is far more versatile and it costs less than half the price!

The sparks can also take on a variety of appearances based on how you set the interrupter. With a low BPS and pulse width, the sparks resemble those of a voltage multiplier or Van De Graaff machine. With ultra-long pulse width, the sparks look almost like fire, or the output from a microwave oven transformer. And by raising the BPS and lowering the pulse width, you can create everything from thick, vacuum-tube coil type bolts, to angry, spark gap coil-style lightning.

Not getting the spark lengths you hoped for? Here are some pro tips for getting the largest sparks possible from your design:

-     Raise the coupling as much as possible without it losing energy to coronal discharge and intercoil arcing

-     Use fewer turns (and therefore a lower impedance). Remember though, 6 ohms is just about the minimum impedance this circuit can handle (unless you get some REALLY powerful IGBTs)

-     Add a topload. This can DOUBLE the output size, especially if you use a smaller breakout point. Make sure to use a breakout point, otherwise the resonant voltage rise will become too great and intercoil arcing will likely result (along with feedback to the primary circuit, and possible transistor damage)

-     Crank up the Variac and boost the voltage from 340VDC to 400VDC

One of the most iconic feature in modern Tesla coils is their ability to play music. This is usually done by applying a music signal in place of the monotonous interrupter signal, causing the sparks’ BPS and pulse width to vary in the same way a speaker does when it plays audio. Although it’s not depicted in the circuit diagram, I added a port on the PCB specifically for custom interrupter signals. If you switch off the main, on-board interrupter (this is necessary; if you don’t, the 555 timer will blow) and apply a 12V squarewave music signal, the coil will come to life playing the corresponding tune! In my accompanying YouTube video, I use an Arduino-switched 12VDC supply to musically interrupt my coil, but the audio quality was pretty bad, since the duty was fairly high (50% or so). For best results, use a lower pulse width signal! One possible candidate for a musical interrupter is the humble MP3 player: simply take the headphone output from the player and use it to switch a transistor, which will in turn switch some 12V power source which will be fed into the PCB’s music port.